Thiyanga Talagala, Rob J Hyndman, George Athanasopoulos
28 June 2017
No free lunch!
If the reasons for differences in performance of various forecasting methods are explored they may be useful in guiding the choice of the best forecasting method (Reid, 1972).
Some methods perform better on some series than others could not be entirely due to chance (Makridakis et al., 1982).
Propose a classification framework which selects forecast models based on features calculated from time series.
Let N be the number of trees to build.
Grow an un-pruned tree based on the bootstrap sample.
Let N be the number of trees to build.
At each node, select m variables at random from the p variables.
Let N be the number of trees to build.
select the best split-point among the m.
Let N be the number of trees to build.
First auto-correlation coefficient
Sum of squares of first 10 auto-correlation coefficients
Number of significant spikes (out of first five) in the ACF
Sum of squares of first 5 partial auto-correlation coefficients
Number of significant spikes (out of first five) in the PACF
Limit the analysis to non-seasonal time series.
Yearly data of M3 competition.
Limit the analysis to non-seasonal time series.
Yearly data of M3 competition.
Develop a random forest classifier to select the most appropriate forecasting model for yearly data of M3 competition.
Limit the analysis to non-seasonal time series.
Yearly data of M3 competition.
Develop a random forest classifier to select the most appropriate forecasting model for yearly data of M3 competition.
All models are selected using the training sets and model evaluation is done by using the test set.
Develop a more comprehensive set of features that are useful in identifying different data generating processes.
Extend the time series collection to non-seasonal data.
Develop a more comprehensive set of features that are useful in identifying different data generating processes.
Extend the time series collection to non-seasonal data.
Test for several large scale real time series data sets.
slides shared online at: https://github.com/thiyangt/ISF-2017-Cairns-Australia